Hydrostatic Test: A Gauge Drop, a Cold Night, and a 175 psi Valve
You are hydrostatically testing a new wet-pipe system in a six-story building. The working pressure at the basement main is 145 psi. The gauge on the basement main reads 200 psi when the test begins; the gauge at the second-floor riser, 55 ft above, reads 176 psi. A 6 in. alarm check valve at the base of the riser carries a 175 psi working-pressure rating. Overnight the ambient temperature in the building falls from 74°F to 52°F. After 30 minutes the basement gauge has fallen to 190 psi while the upper gauge still reads 176 psi, and no water is visible anywhere. The AHJ arrives in the morning.
1. Is the 200 psi test pressure at the basement main acceptable, given the 175 psi alarm valve? What do you do about it?
Read the worked answer
Test pressure is set by the code, but no component may be subjected to a pressure above its listed rating. The 200 psi test pressure itself is correct for a system whose working pressure does not exceed 150 psi, and it must be held for two hours. The problem is the 175 psi alarm valve sitting in that piping at the base of the riser.
The practical resolution is to keep the 175 psi-rated device out of the high-pressure test: temporarily remove it or isolate it with a blind flange or spool so the piping can be tested at 200 psi, then test the valve assembly at its own permitted pressure separately. This is a code-and-AHJ conversation, not something to decide silently on the job - the AHJ witnessing the test is the right person to agree the approach with before the pump is started. What is not acceptable is testing the valve at 200 psi and documenting it as passing.
2. Explain the difference between the two gauge readings, then evaluate the pressure drop.
Read the worked answer
The gauge difference is correct and expected. Static head in a column of water is about 0.433 psi per foot: 55 ft × 0.433 = 23.8 psi, and 200 − 23.8 ≈ 176 psi, matching the upper gauge. Nothing is wrong there - it also confirms both gauges are reading sensibly.
The drop needs diagnosis, not dismissal. A 10 psi loss with no visible water is normally explained by one of two things:
- Thermal contraction: a 22°F drop in a closed, restrained system will contract the water and drop the pressure. This is consistent with the overnight temperature fall and with the pressure stabilizing rather than continuing to fall.
- A leak at a joint, gasket, or the isolated valve blank that is running into a wall cavity, a floor drain, or an underground line and is not visible.
The test method: repressurize to 200 psi, mark the gauge, and observe with the makeup source valved off. A pressure that falls and then holds as the building warms, with no makeup water taken, indicates thermal effects; a steady, continuing loss indicates a leak, and the system is sectioned to find it. Do not simply declare the test passed because the number 'settled.'
3. Where should the test gauge be located, and what else must be true for a valid two-hour test?
Read the worked answer
- Gauge location: at the lowest elevation point in the system, because that is where the test pressure is highest - gauging at the top would allow the bottom of the system to exceed 200 psi. The upper gauge is useful information but is not the governing reading.
- Gauge selection: a range that puts the reading in the middle of the scale, and a current calibration sticker; a 0–600 psi gauge reading 200 psi is too coarse to detect a meaningful loss.
- Duration and stability: 200 psi held for two hours with no drop attributable to leakage. Thermal swings are documented by temperature record and by repressurization, not ignored.
- System completeness: all piping installed, joints exposed for inspection, and any pressure-limiting components isolated.
- Documentation and witness: the AHJ observes the test; the record shows test pressure, gauge locations, start and end readings, duration, temperatures, and the components isolated.
Getting this right the first time is worth far more than re-draining a six-story system after a failed witness test.